State-Space Parameter Estimation With Reduced Initial Value Dependence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing model parameter estimation techniques for storage batteries require setting initial values for all parameters, which can lead to divergent or locally optimal solutions, especially in systems with limited prior information.
Innovation Solution
A model parameter estimation method that separates the estimation into nonlinear and linear parameters, allowing initial values to be set only for nonlinear parameters, using a state space model and iterative updates to achieve convergence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If initial values are set for all parameters in model parameter estimation, then estimation accuracy can be improved, but the complexity of operation increases and the method becomes difficult to apply to systems with little prior information
Solution Approach 1:
The patent segments the parameter estimation process by dividing parameters into two categories: parameters that require initial values and parameters that do not. This segmentation allows the system to maintain high estimation accuracy while reducing the operational burden of setting initial values for all parameters, directly resolving the technical contradiction.
2Measurement precision
If initial values are set based on prior information, then parameter estimation accuracy improves, but the adaptability to systems with little prior information decreases
Solution Approach 1:
By segmenting parameters into those requiring initial values and those not requiring them, the patent enables the method to be applied to systems with little prior information while maintaining estimation accuracy for the parameters that do require initial values, thus improving adaptability.
Solution Approach 2:
The patent creates a universal parameter estimation method that can be applied to different types of systems regardless of the availability of prior information. The method automatically adapts to the system characteristics by identifying which parameters need initial values and which don't, making it versatile across various applications.
3Measurement precision
If all parameters are estimated simultaneously, then comprehensive model accuracy is achieved, but computational complexity and convergence difficulty increase
Solution Approach 1:
The patent simplifies the estimation process by segmenting it into two distinct stages: first estimating parameters without initial values, then estimating parameters with initial values. This segmentation reduces computational complexity and avoids the convergence issues associated with simultaneous estimation of all parameters.
Data Source
AI summary
This model parameter estimation device comprises a state quantity calculator to calculate state quantities indicating a state of a target system with respect to a measured value of an input on the basis of a state equation obtained by assigning values to nonlinear parameters for a model that represents the target system using the nonlinear parameters and linear parameters, and time-series data of the input and an output of the target system, a linear parameter estimator to estimate the linear parameters that minimize an error between the measured value of the output and an estimated value of the output calculated on the basis of the model, the state quantities, and the measured value of the input, and a nonlinear parameter updater to repeatedly update the values of the nonlinear parameters so as to make the minimized error to be small until a convergence condition is satisfied.